New York bans AI chatbots from acting like friends with kids

New York will soon stop AI chatbots from acting like a friend to kids. Gov. Kathy Hochul signed the law on Oct. 28, 2026, and it takes effect Jan. 1, 2027. Companies that break it can be fined up to $25,000 for each violation, and parents will be able to file complaints with the state.

What changed

The law bars companies from offering unsafe chatbot features to minors. It names four: chatbots that pretend to be human, build emotional dependency, encourage self-harm or hold sexually explicit conversations.

Companies must block these features on any account held by someone under 18. They must also block them for any user whose age has not been verified. Providers have to check age using multiple age assurance methods, which means more than one way of confirming how old a user is. Once a user is verified, the age information must be deleted.

How complaints will work

The state attorney general will publish a new website where New Yorkers can file complaints. The material does not say when the site will launch or how complaints will be handled.

Why lawmakers acted

Children often treat chatbots as if they were people. Surveys have shown kids ages 11 to 16 cannot really explain what AI is, and many believe chatbots have human-like feelings and consciousness. The case of Adam Raine, a teenager who died by suicide, is cited as one reason for the law. His conversations with ChatGPT showed it offered to write his suicide note.

A 2026 Common Sense Media report shows how common AI is among kids:

• 86% of kids ages 9 to 17 interact with AI, and about one in four use it daily.
• 57% of kids who use AI have used it for information about their health or body.
• Half have asked AI for advice about their future or goals.
• 37% have used chatbots to talk about feelings or personal problems.

Separate studies have found AI gives misleading or inaccurate health information 50% of the time.

Is it enough?

Experts say no, not alone. Dona Fraser of BBB National Programs, an independent industry accountability group, called the law a step in the right direction. She said companies should not only add guardrails after harm happens but should understand the risks and build around them.

She also said industry self-regulation does not work, pointing to the Meta settlements as an example. In her view, independent third-party accountability is needed, along with companies doing their part.

What happens next

The rules start Jan. 1, 2027. Still unclear are which age checks companies will use and how the state will enforce the law. Until then, parents may want to ask their kids how they use chatbots, especially for health questions and personal worries.